Environ. Sci. Technol. 2007, 41, 3249-3253
The Origin and Significance of Short-Term Variability of Semivolatile Contaminants in Air M A T T H E W M A C L E O D , * ,† MARTIN SCHERINGER,† HEIKE PODEY,† KEVIN C. JONES,‡ AND KONRAD HUNGERBU ¨ HLER† Institute for Chemical and Bioengineering, ETH Zurich, CH-8093, Zurich Switzerland, and Centre for Chemicals Management and Environmental Science Department, Lancaster Environment Centre, Lancaster University, LA1 4YQ, United Kingdom.
Persistent semivolatile contaminants such as polychlorinated biphenyls (PCBs) cycle between air and surface media in the environment. At different locations and times, PCB concentrations in air over a diel (24-hour) period have been observed to have maxima either during the day or at night. These observations have been interpreted as evidence of temperature-mediated air-surface exchange and of degrading reactions with hydroxyl radicals in the atmosphere. However, a general explanation of the processes responsible for the observed diel variability in concentrations has not been provided. Here, we interpret diel monitoring data using a multimedia mass balance model parametrized with local data on temperature, wind speed, atmospheric mixing height, and hydroxyl radical concentrations. We demonstrate that four factors are sufficient to account for the variability of PCB concentrations in air over a diel period; temperature, local atmospheric stability, hydroxyl radical concentration, and source type. We apply the model to re-interpret past diel monitoring studies and find that the observed patterns of concentrations can be rationalized by consideration of these factors. Using insights from this study, future diel monitoring campaigns can be targeted to observe the influence of specific fate and transport processes. Such studies will contribute to more accurate understanding of the processes controlling the shortterm local, and long-term global fate of persistent semivolatile contaminants.
and predict the environmental fate of SVOCs, particularly their transport to, and deposition in, remote ecosystems. However, despite years of study, there remain many unresolved questions about the sources, transport pathways, and ultimate sinks of persistent SVOCs. In the past decade, field studies from different locations have reported concentrations of SVOCs in air measured over diel (24-hour) periods. Results and interpretations of data gathered in these studies have been varied. In some cases, maximum concentrations in air were observed during the day and minimum concentrations at night, coincident with maximum and minimum temperatures. This was interpreted as evidence of temperature-mediated reversible air-surface exchange (3-6). In other cases, maximum concentrations were observed during the night and diel variability was attributed to different possible mechanisms; degrading reactions with hydroxyl radicals in the atmosphere during daylight hours (7, 8), or stable atmospheric conditions during nighttime in areas above probable volatilization sources (9). In still other cases, no regular pattern of variability in concentrations in air was observed over consecutive diel periods (10, 11). Here we use a multimedia mass balance model to demonstrate that source characteristics and three key environmental factors are sufficient to mechanistically explain the patterns of diel variability observed in atmospheric concentrations of persistent SVOCs. The environmental factors are (1) temperature, (2) atmospheric stability (mixing height and surface wind speed), and (3) concentration of photochemically generated hydroxyl radicals. These three factors also explain why no regular pattern is observed in some cases. Concurrent diel cycles in the three controlling environmental factors conspire to affect concentrations of SVOCs in the atmosphere near the surface. Diel cycles in temperature (which affects air-surface partitioning) and incident solar radiation (which photolyzes precursor oxidants into hydroxyl radicals that degrade gas-phase SVOCs) are well known. In addition, during periods when dominant high pressure systems are in place, a stable nocturnal boundary layer that is in direct contact with the surface may develop (12). The nocturnal boundary layer forms slightly before sunset when convective thermal eddies cease because of surface cooling and breaks up shortly after sunrise when surface heating resumes. The nocturnal boundary layer is typically less than 200 meters deep and is characterized by very calm surface winds.
Materials and Methods Introduction Persistent semivolatile organic chemicals (SVOCs) cause large-scale and long-term environmental contamination. During their lifetime in the global environment, SVOCs may undergo several cycles of volatilization to air and deposition to soil, water, and vegetation. This “hopping” is believed to facilitate their overall migration toward the poles (1, 2). The investigation of dynamic exchange of SVOCs between the atmosphere and the global surface with field monitoring campaigns and using models is, therefore, an important research priority. The goal of these studies is to understand * Corresponding author phone: +41 44 632 3171; fax +41 44 632 1189; e-mail:
[email protected]. † ETH Zurich. ‡ Lancaster University. 10.1021/es062135w CCC: $37.00 Published on Web 03/31/2007
2007 American Chemical Society
We have developed a multimedia mass balance model that quantitatively describes the influence of the above three environmental factors on the diel variability of SVOC concentrations in the atmosphere. The diel mass balance model is based on the Berkeley-Trent (BETR) contaminant fate modeling framework, which is described in detail elsewhere (13). BETR is a fugacity-based model that can be run for steady-state (Mackay Level III) or non-steady state (Mackay Level IV) scenarios (14). The general structure of the model is illustrated in Figure 1. The model describes the kinetics of exchange of contaminants between seven environmental compartments; two atmospheric layers, vegetation, soil, freshwater, freshwater sediments, and coastal water. We calculate time-variant contaminant concentrations in all compartments by numerical integration of the constraining mass balance equations. Initial conditions for each model run were set at steadyVOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3249
FIGURE 1. Structure of the regional environment of the diel model and fate and transformation pathways for chemicals. Solid arrows represent intercompartmental transfers, dashed arrows represent advective exchange with the background environment, and lightning bolts represent degrading reactions. state concentrations calculated for median values of temperature, boundary layer height, wind speed, and hydroxyl radical concentration. After initialization, the model was run for 8 days of model time, which was sufficient to allow results to stabilize into a repeating diel pattern. We used Monte Carlo analysis to estimate uncertainty in calculated PCB concentrations from estimated uncertainty and variability in input parameters, as described in the Supporting Information. The formation of a stable nocturnal boundary layer is modeled by blocking exchange between the lower and upper air compartments of the model. Thus the height of the lower air compartment is selected to represent the height of the nocturnal boundary layer, and the height of the upper air compartment represents the height of the mixed convective layer during the daytime. Concurrent with the formation of the boundary layer, wind speed in the lower air compartment is reduced at night. Here, we focus on applying the model to describe diel variability of PCB concentrations in the atmosphere. PCBs are archetypal persistent SVOCs, and the availability of reliable physicochemical property data for the entire range of PCB congeners facilitates application of the model (15, 16). Gas-particle partitioning in the atmosphere is modeled using a relationship with the octanol-air partition coefficient (17). The pseudo first-order degradation rate constant of gasphase PCBs is calculated as the product of hydroxyl radical concentration and the second-order degradation rate constants determined by Anderson and Hites (16). Particle associated PCBs in the atmosphere are assumed to be not degraded by reaction with hydroxyl radicals. Physicochemical properties, degradation half-lives, and associated uncertainty ranges used as input to the model are given in the Supporting Information. The temperature dependence of degradation rate constants in all media and partition coefficients between air, water, and sorbed phases are modeled.
Model Parametrization and Results Figure 2 shows modeled and measured PCB concentrations in air from three field studies (upper two rows of plots), and forcing functions applied as model inputs compared with data on environmental conditions at the study sites or at similar sites (lower three rows of plots). Dates and times on the horizontal axes in Figure 2 refer to local time at the sampling site. Solid lines (s) are (1) modeled PCB concentration in the lower air compartment using default input values (top two rows), and (2) default values of forcing 3250
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007
functions used as input to the model for temperature, mixing height, and hydroxyl radical concentration (bottom three rows). Filled circles (b) are field data gathered during sampling except for mixing height data, which was extracted from the National Center for Environmental Prediction global reanalysis of climate data (18) and temperature data from Chicago, which was obtained from a weather almanac (19). No direct measurements of hydroxyl radical concentrations were available for Hazelrigg or Chicago. Empty circles (O) are median hydroxyl radical concentrations observed in New York City by Ren et al. (20) in July 2001 and are provided for comparison with assumed hydroxyl radical concentrations in Chicago. Hydroxyl radical concentrations at Hazelrigg were estimated based on an assumed 24 hour average global background concentration of 9.7 × 105 molecules/cm3 (16). Dotted lines (‚‚‚) in Figure 2 are 5th and 95th percentile values of calculated PCB concentrations and forcing functions. The PCB congeners and time periods illustrated in Figure 2 were selected as representative of results for all congeners at each site. Complete details of the model parametrization, figures showing the wind speed forcing functions, and figures comparing model results to field data for additional PCB congeners over the entire sampling period at each site are given in the Supporting Information. Hazelrigg, UK, August 19-23, 1995 (3). Monitoring at Hazelrigg, UK in August, 1995 took place during an unusual period of prolonged high atmospheric pressure over Northwest England (3). Hazelrigg is in a rural area, and we therefore modeled the PCB source to the region as a constant background concentration in inflowing air. The magnitude of the background concentration acting as a source to the model region was tuned to achieve the best possible fit with observed data for each congener. The fitted background concentrations agree well with quarterly average PCB concentrations in air measured at the Hazelrigg site during this time period (see the Supporting Information). Observed PCB concentrations in the Hazelrigg study follow a clear diel pattern with maxima during the day and minima at night. This pattern was interpreted by the authors of the original study as being consistent with rapid temperaturemediated air-surface exchange of PCBs and near-equilibrium partitioning between the atmosphere and the surface (3). However, our model analysis indicates that processes associated with boundary layer formation are also important. During the night, when the nocturnal boundary layer is in place, the lower air compartment is isolated from inflow of PCBs from background sources and PCB concentrations fall. This “nocturnal depletion” of PCBs is driven by cooler nighttime temperatures that induce net deposition to the surface, and is enhanced by the relatively small volume of the nocturnal boundary layer. During the day PCB concentrations in the lower air compartment rise slightly above the background concentration. The increase is partially attributable to temperature driven re-volatilization from the surface, as suggested by the authors of the original study (3), but also to downward mixing of PCBs from the upper air compartment. Analogous environmental conditions and source characteristics were present during the studies reported by Hornbuckle et al. (4) and by Wallace and Hites (5), who observed similar diel patterns in concentrations of SVOCs (see details in the Supporting Information). Chicago, July 24-27, 1994 (8). Environmental conditions at Chicago in July 1994 were very similar to those at Hazelrigg, UK in August 1995; however, the qualitative pattern of diel variability in PCB concentrations is reversed with higher concentrations observed at night. The authors of the original study interpreted the lower daytime concentrations as evidence of destruction of gas-phase PCBs by hydroxyl radicals. However, they acknowledge that their interpretation
FIGURE 2. Measured and modeled concentrations of selected PCB congeners in air (top two rows) and environmental conditions (bottom three rows) during diel monitoring campaigns (3, 7, 8). Modeled and measured PCB concentrations at Hazelrigg and Finokalia are the sum of gas phase and particle bound fractions; data for Chicago are for the gas phase only. PCB concentrations are modeled at Hazelrigg and Finokalia assuming a constant background inflow concentration in air and at Chicago assuming a high in-place PCB concentration in soils and uncontaminated inflowing air. implies degradation rate constants that are, on average, five times higher than values derived from laboratory experiments and that other processes might have enhanced the overall rates of disappearance of PCBs during the day (8). Urban areas like Chicago have been shown to be net sources of PCBs to the atmosphere in monitoring studies along urban-to-rural transects (21). Likely PCB sources in urban areas are volatilization from building materials, storage sites, and legacy electrical equipment. These are not explicitly described in the model. We, therefore, modeled the PCB source at Chicago as a large, in-place contamination in soils, and assumed inflowing air to be uncontaminated. Under these model conditions, soil acts as a diffusive source of PCBs and represents all urban sources to the atmosphere. There is net volatilization to the atmosphere over the entire diel period; therefore higher nighttime concentrations are attributable to a smaller volume of air available to dilute the volatilized PCBs compared to daytime. However, the strength of the volatilization source also varies with temperature, with highest source strength during the day. The effects of
temperature and mixing height, therefore, work in opposition to each other under this scenario. For this scenario, the model predicts lower concentrations during the day than at night for all congeners. Model runs including diel variability in temperature, mixing height, or hydroxyl radical concentration alone demonstrate that variability in mixing height dictates the diel trend in PCB concentrations and affects all PCB congeners equally. However, higher chlorinated congeners have higher enthalpies of vaporization and their rate of volatilization is, therefore, increased more strongly by rising temperatures in the morning. More highly chlorinated PCBs thus show weaker diel variability than the lower chlorinated congeners because the daytime volatilization rate of the highly chlorinated congeners is more strongly increased by higher temperatures during the day. Figure 3 shows a comparison of the natural logarithm of observed and modeled ratios of day-to-night concentrations of PCBs with different numbers of chlorine substituents. The model satisfactorily reproduces the trend in day-to-night VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3251
FIGURE 3. Natural logarithm of observed (b) and modeled (s) ratio of day-to-night concentration of PCB congeners in the gas phase measured at Chicago, July 24-27, 1994 (8). Model results are shown as a central value with 5th and 95th percentile values calculated from the Monte Carlo uncertainty analysis. concentration ratios observed at Chicago for PCBs with 2 to 6 chlorines. Model runs with varying and with constant hydroxyl radical concentrations show that the trend in dayto-night concentration ratios is not attributable to degradation by hydroxyl radicals. Analogous environmental conditions and source characteristics were present during the study by Bidleman et al. (9) who reported day-night concentration ratios of toxaphene congeners in air above contaminated agricultural soils, as described in the Supporting Information. Finokalia, Greece, August 19-23, 2001 (7). Environmental conditions at Finokalia, Greece in August 2001 are unique among diel monitoring studies of persistent SVOCs carried out to date (7). Atmospheric mixing height during sampling was low and very nearly constant, and temperature varied only within a narrow range (Figure 2). In addition, hydroxyl radical concentration, which was monitored in realtime concurrently with PCB sample collection, was extremely high, approximately 7 times higher than concentrations typical of 24-hour global averages assumed at Hazelrigg and 3.5 times higher than typical urban concentrations assumed at Chicago. Finokalia is in a remote location on the northern coast of the Isle of Crete, and therefore, we modeled PCB sources to the site as a constant background concentration in inflowing air. The model predicts that, under these unusual environmental conditions, reaction lifetimes during the day for PCB congeners with five or fewer chlorine substitutents are sufficiently short that daytime depletion due to degrading reactions with hydroxyl radicals could be observable. For these congeners, concurrent variability in PCB concentrations due to variable windspeed and volatilization from the surface is not strong enough to completely obscure the daytime depletion due to degrading reactions. The additional variability in the field data between August 21 and 22 that is not reproduced by the model could be caused by fluctuation of the inflow concentrations that is covariant across congeners, whereas the model assumes a constant inflow concentration of PCBs in air. In this case, our modeling supports the interpretation presented by the authors of the original study; the pattern of diel variability they observed was attributable to hydroxyl radical degradation of gas-phase PCBs. We are not aware of any analogous studies showing the direct influence of hydroxyl radical degradation on PCB concentrations measured in the field. Other Diel Monitoring Studies. In some studies of diel variability of persistent SVOC concentrations in air reported in the literature, no clear and repeating pattern of concen3252
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007
FIGURE 4. Temperature (6), atmospheric mixing height (18), and concentration of PCB28 in whole air (6) measured at Pigeon Lake, Canada in April, 2000. trations was observed. One example is the study by Gouin et al. (6), who reported diel variability in polychlorinated biphenyls and polybrominated diphenyl ethers (PBDEs) in air at Pigeon Lake, Canada in April, 2000. The monitoring site is remote from probable sources of PCBs and PBDEs, and therefore, the source of contamination in the area is expected to be transport in air. Temperature followed a repeating pattern of diel variability over the sampling period, but was trending slightly downward. The source conditions and temperature were, therefore, similar to those at Hazelrigg, UK; however, Gouin et al. did not observe a consistent relationship between PCB and PBDE concentrations and temperature. Figure 4 shows variability in temperature and mixing height, and monitoring results for PCB 28 that are representative of results for other PCB congeners and PBDEs in this study. As shown in Figure 4, the atmospheric mixing height at Pigeon Lake on April 26, 2000 rose to nearly two times the maximum height on the preceding 3 days. Prior to this event, the air above 750 meters had been out of contact with the surface for several days, and may have become depleted in PCBs. Therefore, the lack of a significant rise in PCB concentrations with temperature during the day on the 26th is likely due to uncontaminated air mixing down from the upper levels of the atmosphere. We explored this hypothesis using the diel model and estimated that the amplitude of the peak in PCB concentrations during the day would be reduced by about a factor of 2 by such an event, which is consistent with the observations reported by Gouin et al. We have collected environmental data for two other studies where clear diel patterns in SVOC concentrations in air were not observed (10, 11) and found that monitoring took place during a period when mixing height did not follow a stable and repeating pattern (see the Supporting Information). This again shows the importance of mixing height in combination with temperature as key factors controlling the diel variability of SVOC concentrations. Jaward et al. (22) reported diel variability in PCB concentrations in air over the South Atlantic Ocean. They
observed higher PCB concentrations during the day than at night during periods of constant temperature and atmospheric mixing height. Thus, these data are anomalous when compared to diel monitoring studies at land-based sites and indicate that different fate processes may be operative near the air-ocean interface. In future, we plan to adapt our mass balance model to examine different hypotheses to explain these data, including the possible influence of dynamic partitioning to phytoplankton near the surface of the ocean (22).
Recommendations Our analysis using a mass balance chemical fate model has identified four controlling factors that determine the shortterm variability of concentrations of persistent SVOCs in air: (1) source characteristics, (2) temperature dependent airsurface partitioning, (3) atmospheric mixing height and stability, and (4) hydroxyl radical concentration. Regular and interpretable patterns of diel variability in observed SVOC concentrations in air can only be expected when environmental conditions are also stable and repeating. Future diel monitoring campaigns should be designed with careful consideration of these four controlling factors. Selecting sampling sites and times that maximize the chances of observing a particular aspect of SVOC fate and interpretation of monitoring data with a mass balance model will allow more accurate characterization of environmental fate processes. The quantitative understanding of processes that control both the short-term local and long-term global concentrations of persistent semivolatile contaminants derived from such studies will support more effective management of chemical products. Identification of sites contaminated by persistent organic pollutants that may act as ongoing sources to the global environment is a requirement under Article 6 of the Stockholm Convention on persistent organic pollutants (23). As demonstrated here, combined field and modeling studies provide the means to identify ongoing sources from contaminated areas and discriminate them from secondary sources due to recycling from water, soils, and sediments that were contaminated by long-range transport and deposition. Thus the scientific tools described here can be used to support implementation of national obligations under the Stockholm Convention.
Acknowledgments We thank R. G. M Lee, C. Halsall, L. Totten, and E. Stephanou for providing raw data from their monitoring studies.
Supporting Information Available Details of parametrization of the model and the Monte Carlo uncertainty analysis. Model results and discussion of additional diel monitoring data for the three case studies shown in Figure 2 and other studies. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Wania, F.; Mackay, D. Global fractionation and cold condensation of low volatility organochlorine compounds in polar regions. Ambio 1993, 22, 10-18. (2) Gouin, T.; Mackay, D.; Jones, K. C.; Harner, T.; Meijer, S. N. Evidence for the “grasshopper” effect and fractionation during long-range atmospheric transport of organic contaminants. Environ. Pollut. 2004, 128, 139-148. (3) Lee, R. G. M.; Hung, H.; Mackay, D.; Jones, K. C. Measurement and modeling of the diurnal cycling of atmospheric PCBs and PAHs. Environ. Sci. Technol. 1998, 32, 2172-2179.
(4) Hornbuckle, K. C.; Eisenreich, S. J. Dynamics of gaseous semivolatile organic compounds in a terrestrial ecosystemEffects of diurnal and seasonal climate variations. Atmos. Environ. 1996, 30, 3935-3945. (5) Wallace, J. C.; Hites, R. A. Diurnal variations in atmospheric concentrations of polychlorinated biphenyls and endosulfan: Implications for sampling protocols. Environ. Sci. Technol. 1996, 30, 444-446. (6) Gouin, T.; Thomas, G. O.; Cousins, I.; Barber, J.; Mackay, D.; Jones, K. C. Air-surface exchange of polybrominated biphenyl ethers and polychlorinated biphenyls. Environ. Sci. Technol. 2002, 36, 1426-1434. (7) Mandalakis, M.; Berresheim, H.; Stephanou, E. G. Direct evidence for destruction of polychlorobiphenyls by OH radicals in the subtropical troposphere. Environ. Sci. Technol. 2003, 37, 542-547. (8) Totten, L. A.; Eisenreich, S. J.; Brunciak, P. A. Evidence for destruction of PCBs by the OH radical in urban atmospheres. Chemosphere 2002, 47, 735-746. (9) Bidleman, T. E.; Leone, A. Soil-air relationships for toxaphene in the southern United States. Environ. Toxicol. Chem. 2004, 23, 2337-2342. (10) Sofuoglu, A.; Cetin, E.; Bozacioglu, S. S.; Sener, G. D.; Odabasi, M. Short-term variation in ambient concentrations and gas/ particle partitioning of organochlorine pesticides in Izmir, Turkey. Atmos. Environ. 2004, 38, 4483-4493. (11) Lohmann, R.; Brunciak, P. A.; Dachs, J.; Gigliotti, C. L.; Nelson, E.; Van Ry, D.; Glenn, T.; Eisenreich, S. J.; Jones, J. L.; Jones, K. C. Processes controlling diurnal variations of PCDD/Fs in the New Jersey coastal atmosphere. Atmos. Environ. 2003, 37, 959969. (12) Stull, R. B. An Introduction to Boundary Layer Meteorology; Kluwar Academic Publishers: Boston, MA, 1988. (13) MacLeod, M.; Woodfine, D. G.; Mackay, D.; McKone, T.; Bennett, D.; Maddalena, R. BETR North America: A regionally segmented multimedia contaminant fate model for North America. Environ. Sci. Pollut. Res. 2001, 8, 156-163. (14) Mackay, D. Multimedia Environmental Models: The Fugacity Approach; Lewis Publishers: Boca Raton, FL, 2001. (15) Schenker, U.; MacLeod, M.; Scheringer, M.; Hungerbu ¨ hler, K. Improving data quality for environmental fate models: A leastsquares adjustment procedure for harmonizing physicochemical properties of organic compounds. Environ. Sci. Technol. 2005, 39, 8434-8441. (16) Anderson, P. N.; Hites, R. A. OH radical reactions: The major removal pathway for polychlorinated biphenyls from the atmosphere. Environ. Sci. Technol. 1996, 30, 1756-1763. (17) Harner, T.; Bidleman, T. F. Octanol-air partition coefficient for describing particle/gas partitioning of aromatic compounds in urban air. Environ. Sci. Technol. 1998, 32, 1494-1502. (18) NCEP. National Center for Environmental Prediction Global Reanalysis of Climate Data; 2005; http://www.cpc.ncep.noaa.gov/ products/wesley/reanalysis.html. (19) The Old Farmer’s Almanac; Yankee Publishing Inc.: Dublin, New Hampshire, 2006; http://www.almanac.com/. (20) Ren, X. R.; Harder, H.; Martinez, M.; Lesher, R. L.; Oliger, A.; Simpas, J. B.; Brune, W. H.; Schwab, J. J.; Demerjian, K. L.; He, Y.; Zhou, X. L.; Gao, H. G. OH and HO2 chemistry in the urban atmosphere of New York City. Atmos. Environ. 2003, 37, 36393651. (21) Harner, T.; Shoeib, M.; Diamond, M.; Stern, G.; Rosenberg, B. Using passive air samplers to assess urban-rural, trends for persistent organic pollutants. 1. Polychlorinated biphenyls and organochlorine pesticides. Environ. Sci. Technol. 2004, 38, 44744483. (22) Jaward, F. M.; Barber, J. L.; Booij, K.; Dachs, J.; Lohmann, R.; Jones, K. C. Evidence for dynamic air-water coupling and cycling of persistent organic pollutants over the open Atlantic Ocean. Environ. Sci. Technol. 2004, 38, 2617-2625. (23) Stockholm Convention on Persistent Organic Pollutants; 2005; http://www.pops.int/.
Received for review September 7, 2006. Revised manuscript received February 22, 2007. Accepted March 5, 2007. ES062135W
VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3253
